Investigating the structure of protein similarity networks both on sequence and structural level

The current work presents a study concerning similarity networks for a well defined and used dataset of proteins, constructed using sequence and structural derived similarity criteria. The analysis is made on the initial set of proteins and the subsets of proteins that yield to fully connected networks using both similarity criteria. Several parameters describing the networks are reported, whereas the values of clustering coefficient and characteristic path length classify both types of fully connected networks into the class of small-world networks. This reveals the grouping of proteins in clusters and the existence of random edges in the networks due to similarity transition among proteins.